Unicon documentation migration is in progress. You might find some broken links or experience minor issues in the documentation. We are working on resolving these issues.

X

Defining a new cluster instance (earlier versions)

– for [SCG 1 2201] and earlier versions –

For versions earlier than SCG 1 2209, new instances must be explicitly defined as a cluster instance in the configuration of the WebAdmin interface: When you add a new instance, no automatic clustering takes place - new random values are generated for each newly installed instance and the administrator must apply cluster key and database password. For information on the procedure from SCG 1 2209, see Adding a new gateway instance.

Note:

Before you change a cluster key, make sure that at least one other instance uses the new cluster key.

  1. Install another gateway instance and start the machine.For further information, see Installation.

  2. Allow the system two to three minutes before you log on to the SCG WebAdmin interface.

  3. Switch to the Configuration > SCG cluster view and edit the following options:

    Option Description
    Database password Copy and paste the database password from another instance.
    Cluster key Copy and paste the cluster key from this other instance.

    Clustering is only done if both instances have the same hash value. The hash value is built from the database password and the cluster key.

  4. Leave the WebAdmin interface of your new instance.

The new instance is added to the cluster. Once the instance is running with the new configuration, it is shown in the WebAdmin interface of the main instance, in the Gateways view.

When a gateway instance is started, a broadcast message is sent, other instances with the same hash value can react. The instances communicate with each other via TCP port 27017 within their network.

Note:

Make sure that the instances can reach the Scout Server via appropriate network routes.

Do not change database password and cluster key after you have created a cluster. Changing any of the values may result in the instance becoming inoperable.

Defining a new cluster instance (earlier versions)